Most investors, when discussing technology, are actually talking about two different things: software, hardware and internet companies as a distinct industry, and technology as a general-purpose capability reshaping the cost structure and competitive dynamics of other industries. The two are frequently conflated, distorting judgments about where value actually forms. Havrion Capital analyzes them separately, evaluating technology companies on their own business models and separately evaluating how technology reshapes competitive structure inside other industries, since the latter often determines whether the former can sustain its economics.
This distinction is not academic. A technology company's valuation logic depends on whether it is selling efficiency gains to other industries — demand that is stable but comes with limited pricing power — or building an independent new industry of its own, an opportunity that is rarer and carries materially more uncertainty.
Industry-specific infrastructure
Provides underlying capability to a specific industry (industrial quality-inspection algorithms, enterprise data platforms). Switching costs are high and customer relationships run deep, making this one of the more durable pockets of value in the sector.
General-purpose infrastructure
Compute, storage, networking and development platforms that cut across industries. Scale effects are pronounced, but the competitive field is often already led by a small number of participants, leaving limited room for new entrants.
Vertical applications
Software and tools built for a specific industry's workflow. Value depends on whether the product embeds itself in a customer's core process; applications that remain peripheral tools are readily displaced.
General consumer and utility applications
Mass-market application software, where user-acquisition cost and popularity cycles swing widely. This is the area within the sector that most requires discipline against chasing a passing trend.
Top-left: Industry-specific infrastructure; Top-right: General-purpose infrastructure; Bottom-left: Vertical applications; Bottom-right: General consumer and utility applications
Illustrative framework for organizing judgment, not a quantitative score
Where durable value settles
Three questions determine whether value in a technology investment is likely to endure.
Infrastructure or application
Infrastructure-layer value comes from being relied upon by many applications, with replacement cost rising as dependence deepens. Application-layer value depends more on continuous iteration and customer-relationship maintenance, and is more exposed to competitive displacement. Judging which layer a company occupies matters more than judging how fast it grows.
The real source of switching cost
Switching cost may come from accumulated data, embedded workflows or operator habit, or it may be superficial stickiness created only by contract terms. The former reinforces itself over time; the latter dissolves once a contract lapses. The two are difficult to tell apart in the short run.
Ownership of the distribution channel
A company that owns direct customer relationships retains more pricing power than one that reaches customers through a platform or distributor. The degree of control over distribution often predicts long-run margin better than the technical sophistication of the product itself.
Discipline against fashion cycles
In technology, narratives refresh far faster than business models actually evolve, which requires the evaluation process to keep a deliberate distance from prevailing sentiment.
Separating terminology from economics
How popular a new term becomes has no bearing on the cash-flow characteristics behind it. Evaluation always starts from the latter.
Independent judgment on valuation multiples
A rise in comparable transaction prices is not, by itself, a reason to invest. The assumptions underlying that price require separate verification.
Patience with the exit window
Liquidity in a fashionable niche tends to concentrate in a narrow window. Investment pace does not relax diligence standards simply because that window is approaching.
Competitive dynamics and capital intensity
Competitive structure varies sharply within the sector. At the infrastructure layer, barriers, once established, tend to be relatively stable, and capital needs concentrate in the early build-out phase. At the application layer, competition is continuous and capital needs spread across the life cycle. This divergence means capital intensity cannot be discussed as a single sector-wide characteristic; it has to be measured against the specific value layer in question.
The Havrion Capital Perspective
Within technology, Havrion Capital prioritizes companies where value settles at the infrastructure layer or in applications deeply embedded in an industry's workflow, since these economics align more closely with the long-horizon holding model the platform prefers. Correspondingly, the sector actively avoids areas where value derives mainly from narrative momentum and has not yet formed verifiable switching costs — not because the underlying technology lacks merit, but because its cash-flow predictability does not support the logic of long-term capital allocation.
Growth-stage technology transactions fall mainly under the technology investments strategy. Analysis of technology as a horizontal force — how it reshapes competitive structure in other industries — runs through several of the other industry pages rather than being repeated here.
Related Portfolio Companies
Related Insights
From Project to Product: The Structural Shift in China's Enterprise Software Market
China's enterprise software market is going through a change in economic model — from custom project work billed as a one-time license, toward standardized product sold on subscription. The idea is simple to state and full of traps in execution: most companies that claim to have made the shift still carry a heavy shadow of project economics inside their revenue mix. This piece examines why mid-market vertical depth travels further than broad horizontal ambition, why retention is the one number that does not lie, and where this shift is furthest along, and where it has stalled.
8 min readWhere AI Adoption Is Real in the Industrial Economy
Most industrial AI deployments remain pilots. A smaller set has moved into repeatable, cost-accountable operation. What separates the two has less to do with model sophistication than with data access, deployment capability and ownership of the workflow — the three factors that determine whether a vendor compounds or stalls at the next budget review.
7 min read